CAMBRIDGE, Mass.--To reinvent the fuel business, engineers at biofuel start-up Joule Unlimited tinker with tiny life forms all day.

The four-year-old start-up is on the front lines of a branch of biotechnology that taps into the wealth of knowledge from genome sequencing and powerful computer tools to start from scratch and ask: if you wanted the ideal fuel, how would you make it

The answer they've come up with is a diesel secreted by a genetically engineered microbe in flat plastic bioreactors. The only inputs for its "biofactory" organism are sunlight, pumped-in carbon dioxide, and some nutrients.

Joule Unlimited has a long way to go before it's a commercial home run. But a look at its business strategy and labs here demonstrate the possibilities of biotechnology in reshaping the liquid fuels industry.

After disappointing progress in making ethanol from agriculture wastes or grasses, much of the public attention has turned to plug-in electric vehicles to make transportation greener. But replacing petrofuels and chemicals with plant-based sources is still very much part of the picture.

Joule Unlimited was co-founded by venture capitalists at Flagship Ventures who took a "blue sky" approach to making biofuels at scale after investing in two other biofuels start-ups also using techniques from synthetic biology--Mascoma and LS9.

Rather than use algae or shuffle the genes of industrial workhorses, such as e.coli bacteria, Joule is doing metabolic engineering of cyanobacteria or blue-green algae, which is thought to have evolved 2.9 billion years ago and is the granddaddy of all water-splitting photosynthetic organisms.

During a meeting at the company's offices, in what Cambridge calls "Life Sciences Square," Joule Unlimited's senior vice president of biological services (and the company's first employee), Dan Robertson, shows me a thick Department of Energy report that identifies a number of barriers to making fuel with algae.

The bio-engineers at Joule set out to address all of those barriers, Robertson said. Instead of growing biomass and then extracting the fuel from it, Joule and others want to make fuels secreted directly from micro-organisms. Instead of feeding sugar--typically from sugar cane--to e.coli and then fermenting the solution to brew alcohol-based fuel, Joule has designed a system of continuous fuel production. Algae wasn't pursued because so much water needs to be removed and it's harder to engineer.

"In one place, the light is being absorbed and the carbon dioxide taken up where you can get carbon molecules to ultimately make your product," Robertson explained.

The cyanobacteria in bioreactors produce hydrocarbons, which are siphoned off from the green-colored water solution. There's no feedstock to procure and ship or even biomass to gather and then process. The company is running a pilot facility in Texas, where it is trucking in carbon dioxide for testing purposes. For future sites, it intends to get CO2 from an industrial partner. Emissions from a coal-fired power plant could be fed to the cyanobacteria after normal scrubbing of mercury and arsenic, Robertson said.

Bug control Breeding specific organisms, such as e.coli, for specific purposes has been going on for decades. But advances in biotech are allowing biologists to manipulate organisms with far finer control and speed.

In Joule's case, biologists are optimizing the metabolism of cyanobacteria for their purposes. Instead of taking food and sunlight to make more of themselves, the natural pathways of Joule's genetically modified organisms have been altered to produce alkanes, a hydrocarbon to be mixed with diesel fuel.

They have even been programmed with a "carbon switch" to shift their metabolisms from making more of themselves to making fuel. Joule envisions starting up production of its microbes in its bioreactors and, once they reach a certain density in water, operators will add ingredients to get the organism's metabolism to start pumping out alkanes.

"You essentially co-opt the carbon that would be used for something else (to make fuel) and make the organism feel that it's OK," said Robertson. "It's all very controlling."

In the lab, engineers re-create real-world conditions to isolate different strains by tweaking carbon dioxide, sunlight, and nitrogen levels. For example, one specialized machine allows an engineer to simulate the sunlight conditions for a whole day, representing different seasons and changing temperatures.

With each test, engineers get closer to the optimal gene combinations for different conditions. Since the bioreactors would be placed in different locations, the company has settled on somewhere between 10 and 20 strains, after constructing some 4,000 strains, according to Robertson.

Biologists spend more of their time designing tests or isolating DNA on gene databases, rather than actually performing lab tests, Robertson said. "Because so much genome sequencing has already taken place, there's an immense database available arranged by the chemistry that they do," he said. "There are lots of tools at our disposal, so we can rapidly test things."

Long road to pipelines Joule Unlimited is just one of many companies in the race for a better biofuel technology. The Department of Energy's ARPA-E agency is funding many research efforts in "electrofuels," which make fuels directly from electricity, sunlight, and water. There are dozens of other companies pursuing completely different approaches, such as thermochemical processes.

Analysts at Lux Research call synthetic biology is the "flashiest" technology vying for the lead. Joule Unlimited has high potential, but it still has to prove that it can move from a small-scale operation to commercial scale, Lux Research said. Specifically, it lacks partners and it will require a lot of capital and land to reach cost parity with today's petrofuels, it said in a recent report.

The company's next step is to build a 10-acre demonstration plant located at an industrial facility, such as a power plant or waste water facility, which it expects to do by mid-2012. Then it plans to start construction of a facility larger than 1,000 acres at the end of 2013 that would be able to make 12,000 gallons of fuel per acre per year. When fully scaled up, it would make 15 million gallons a year of fuel.

The cyanobacteria need nutrients, including nitrogen and trace amounts of minerals, to be fed into its bioreactors. And to be commercially successful, its bioreactors need to be engineered for low operational costs and maximum production. Next door to its labs, engineers are designing the actual hardware for growing the fuel, another key part to whether the company will be able to scale up.

Although Joule's employees clearly see the power of metabolic engineering, they are very cognizant of concerns over genetically modified organisms, said Robertson. Its bioreactors will run for six to eight weeks' then the solution is flushed out, the actual organisms are burned, the bioreactors are sterilized, and a fresh medium is put in, he said.

Even with Joule's impressive technical achievement and three patents, it still has a ways to go before it will make a dent in reducing fossil fuel use and pollution. But if the biotechnology techniques it and others are pursuing bear fruit, the face of fuels in the future may well be a petri dish, rather than an oil well.

Updated at 1:15 p.m. PT with corrections. The bioreactors will be made from plastic, not glass, and Joule's demo facility is using trucked-in CO2.